It was a 6N23P. TubeCad doesn't have a model for that tube. Not interested in polemics, enjoy your preamp. I didn't enjoy my headphone amp, that's for sure.
I was expecting this, now I have to explain myself.I built something along these lines:
http://www.diyaudio.com/forums/attachment.php?s=&postid=84773&stamp=1037816258
And a headphone amp with these values, you may say the output impedance was too high and that may be the cause but I did try with a matching transformer too. Same disgusting sound.
As for the reason why they sucked I can't say. My body had a natural rejection against both designs.
Thanks, the detail is super helpful to understand your use case.
A low mu triode isn't a common candidate for SRPP due to the high grid swing requirement. Not that it won't work but I think it may be a little outside the common design envelope.
The headphone amp at 11 mA idle current only produces about 33 milliwatts max into 600 ohms before clipping. I don't know your headphone sensitivity but it sounds a little low.
To me it's a little hard to make these conclusions without some measurements to qualify the sonic impression. I can think of a couple of things e.g. above that could go wrong. If it's that obvious, it should be easily measurable.
Personally I would not try to use a balanced SRPP to drive a reactive variable impedance load. I have the anti triode for that. But it helps to see what people have tried and what didn't work out.
Could you point me towards a good description of an anti-triode?
I am considering an SRPP/SRPP+ driver but if an anti-triode is similar...
Anti triode is a name someone made up to describe a class A push-pull circuit consisting of a triode and an opposing device (triode, pentode, transistor...) such that the opposing device is controlled by the triode in a way that it exactly mirrors the triode signal current and as such does not cancel the 2nd harmonic distortion of the triode. The anti triode stage has the characteristic of a SE triode stage.
Broskie's SRPP+ is essentially the same as the totem pole version of the anti triode. The top device is controlled by the triode as in SRPP and mu-follower circuits, but through a pair of current sensing resistors with the output connected between them. Given high enough gm of the top device, and adjusting the resistor ratio for the finite gm of the top device, the signal current swing will be exactly opposite between the triode and the top device, resulting in push-pull anti-triode operation.
The output impedance will be approximately 1/2 the plate resistance of the triode. The current swing will be 2X the idle current. e.g. for an idle current of 20 mA, the current swing will be +40 mA, -40 mA into the load.
I've used MOSFETs for the opposite device in all of my experiments, with excellent results. I'm not sure using a similar triode for the opposite or top device would preserve the SE response as well.
Last edited:
2nd on Michael's comments. The totem pole version of the anti-triode is the same as the SRPP+, except it goes one step further, using dissimilar gm's to preserve the 2nd harmonic. One could look at it as a fine tuning of an impedance multiplier circuit to preserve the harmonic structure as well. A low capacitance depletion mode Mosfet up top should give the best results.
The anti-triode scheme covers several other P-P configurations besides the totem pole. There is a convenient LTP version (CCS tail with a Mosfet or high gm device on one side), then there is 1/2 of the WE Harmonic Neutralizer (OT B+ current sense (CT'd OT) feedback to just one P-P grid side) should work, and various Op. Amp control schemes too. One could even take the totem pole version and split it between primary and secondary (tube on primary, Mosfet on secondary, essentially a split Circlotron), to lower the B+ by about half (but some complications to fix the small output offset from secondary winding resistance).
The anti-triode scheme allows one to get SE sound with P-P grunt (class A level P-P power, low output Z, good bandwidth) while using an ordinary P-P OT. That's what drove its development. It's also easy to make it continuously adjustable between SE and P-P sound, can adjust to suit the music.
The anti-triode scheme covers several other P-P configurations besides the totem pole. There is a convenient LTP version (CCS tail with a Mosfet or high gm device on one side), then there is 1/2 of the WE Harmonic Neutralizer (OT B+ current sense (CT'd OT) feedback to just one P-P grid side) should work, and various Op. Amp control schemes too. One could even take the totem pole version and split it between primary and secondary (tube on primary, Mosfet on secondary, essentially a split Circlotron), to lower the B+ by about half (but some complications to fix the small output offset from secondary winding resistance).
The anti-triode scheme allows one to get SE sound with P-P grunt (class A level P-P power, low output Z, good bandwidth) while using an ordinary P-P OT. That's what drove its development. It's also easy to make it continuously adjustable between SE and P-P sound, can adjust to suit the music.
Last edited:
Thanks for the explanations - very interesting.
I wonder if an anti-triode or SRPP+ parafeed amp would be a viable mod. to a PP amp? With double the voltage and parafeeding right across the whole primary I wonder if this scheme would approach the power of the original PP?
The purpose of doing this (perhaps rather odd mod) would be to get rid of crossover distortion and to move the reference at the primary from B+ back to ground - hence making the whole feedback-referenced-to-ground thing possible.
I wonder if an anti-triode or SRPP+ parafeed amp would be a viable mod. to a PP amp? With double the voltage and parafeeding right across the whole primary I wonder if this scheme would approach the power of the original PP?
The purpose of doing this (perhaps rather odd mod) would be to get rid of crossover distortion and to move the reference at the primary from B+ back to ground - hence making the whole feedback-referenced-to-ground thing possible.
Never.
What is the point in bridging outputs in tube amps? To get hemorroidal problems with transformer primary impedance? To add dirt from phase splitter?
I use in my last Nirvana project anti-triode SRPP made with pentode (local feedback to 1'st grid) and Semisouth power JFET.
Bridging is fine for SS amps when voltage is limited by used devices, or directly powered from some battery like in pocket players.
Member
Joined 2009
Paid Member
I agree.
Looking at this Ju-Jutsu further (what the heck is Ju-Jutsu ? did you mean Ju-Jitsu?) I don't like it - dangerously high voltages for the sake of using all-tube. In today's world this makes no sense anymore. And for 'purists' it still puts capacitors in the signal path.
Looking at this Ju-Jutsu further (what the heck is Ju-Jutsu ? did you mean Ju-Jitsu?) I don't like it - dangerously high voltages for the sake of using all-tube. In today's world this makes no sense anymore. And for 'purists' it still puts capacitors in the signal path.
As Wavebourn and Bigun said, you have to use uncomfortably high B+ voltages with the totem pole, and the full winding primary impedance will be 4X what the P-P tubes saw before.
With a P-P OT in hand, I would try it with the LTP version of anti-triode. Same B+, you can use the same output tube, but in triode mode or Schaded local feedback mode. The other tube gets changed to a HV N Mosfet. The tail gets a CCS. The CCS will remove the common mode B+ power supply noise from the output in class A mode.
It will suffer sudden clipping in class A, but so does the totem pole. Some small DC in the OT balance issue can develop at high amplitude signals (due to 2nd harmonic dist from the tube), which is fixable by either biasing at a little lower current on the tube (than the Mosfet), or by using a double CCS tail (under each output device) with a cap between.
With a P-P OT in hand, I would try it with the LTP version of anti-triode. Same B+, you can use the same output tube, but in triode mode or Schaded local feedback mode. The other tube gets changed to a HV N Mosfet. The tail gets a CCS. The CCS will remove the common mode B+ power supply noise from the output in class A mode.
It will suffer sudden clipping in class A, but so does the totem pole. Some small DC in the OT balance issue can develop at high amplitude signals (due to 2nd harmonic dist from the tube), which is fixable by either biasing at a little lower current on the tube (than the Mosfet), or by using a double CCS tail (under each output device) with a cap between.
Last edited:
Guys, can we split the anti-triode and related stuff into a new thread if you feel that that would be useful? Post a suggested point to separate the thread if you feel so inclined. Let me know if it isn't and continue here. Thanks! 😀
Member
Joined 2009
Paid Member
The way I see it, the topology that is the subject of this thread, is one in which the plate load for the amplifying (bottom) devices has an active load. Whether this active load is a tube or an anti-triode / CCS / Gyrator is not important other than it must provide excellent PSRR to keep the supply noise out of the primary of the OPT, and that this active load allows the amplifying (bottom) devices to operate in a linear fashion for best results. Lower voltages can compromise the operation of tubes, so using SS on top allows much more comfortable voltages. And arguably, SS on top will outperform a tube on top (technically at least - I'm not commenting on the subjective). There are SS parts available which have sufficient voltage and current rating; they do not drift with age and require no heaters (which would otherwise have to be elevated in voltage too).
So it build down to a PP parafeed amplifier and so attracts the same benefits and disadvantages of paraded outputs.
So from my perspective a discussion about SS on top is relevant for this thread, but the original poster may feel otherwise.
So it build down to a PP parafeed amplifier and so attracts the same benefits and disadvantages of paraded outputs.
So from my perspective a discussion about SS on top is relevant for this thread, but the original poster may feel otherwise.
Last edited:
My appologies to the original poster. I just mentioned the anti-triode as a logical progression from the original PSRR goals, due to the similarity and improved power efficiency. Some of us like to explore creative diversions from the main topic when they seem warranted or useful, apparently the original poster had different ideas. Back to the original programming.
For the Mu follower approach, it seems there are some issues that should be clarified by the OP. The triode (Rp) on top would appear to compromise the PSRR, which was posited as the main problem to solve I believe. Also, the lack of symmetrical current ability for a P-P stage using the Mu follower is curious. Push-push amplifiers would seem to have some limitations.
In the interest of reducing the extra high B+ required for the Mu follower totem pole configuration, I would suggest that the current loading stage -could- be moved over to the secondary side using a LV Mosfet CCS and the primary/tube side ultra-pathed for PSRR. Just a practical suggestion. Another option would be a tertiary winding for the LV CCS. I once added a 3rd winding onto an Edcor GXSE15 and removed the SE gap for such a scheme.
For the Mu follower approach, it seems there are some issues that should be clarified by the OP. The triode (Rp) on top would appear to compromise the PSRR, which was posited as the main problem to solve I believe. Also, the lack of symmetrical current ability for a P-P stage using the Mu follower is curious. Push-push amplifiers would seem to have some limitations.
In the interest of reducing the extra high B+ required for the Mu follower totem pole configuration, I would suggest that the current loading stage -could- be moved over to the secondary side using a LV Mosfet CCS and the primary/tube side ultra-pathed for PSRR. Just a practical suggestion. Another option would be a tertiary winding for the LV CCS. I once added a 3rd winding onto an Edcor GXSE15 and removed the SE gap for such a scheme.
Last edited:
The theory of this design
and explanation of its operation was amply and clearly covered.
Try reading the thread.
Thanks for your contribution.
Last edited:
I think this is substantially correct.
I prefer an all-tube circuit,
in part because of reliability questions as to SS in high-voltage systems.
The elimination of a floating HV heater circuit would require the top-tubes in both stages to be replaced by SS.
If a reliable SS CCS top can be designed with equivalent performance,
then it could be a real advantage.
Discrete SS circuits are not my area of expertise,
nor do I have reliable information on the performance of SS devices in HV circuits.
So I won't be designing a Hybrid version of my amp.
No, I think sincere contributions and alternate circuits such as you propose,
are quite relevant to the thread.
But I would like to see some actual circuit designs,
with some discussion of reliability and performance of key devices.
Last edited:
Well, I'm very comfortable with the high B+ confined to a separate PS section or chassis, as for instance with the top tubes treated as part of the PS.
I don't see the disadvantage of higher impedance loads on the output tubes. Instead this increases linearity.
Putting another CCS on the bottom of the output tubes seems to actually add more complexity to the circuit,
if you've already replaced the top tubes with Mosfets.
I suppose you could eliminate tubes in the output stage altogether,
but this seems to substitute an entirely different ethos,
not one based on tubes and their potentially incredible performance.
As I said, in A/B tests, my tube amp blew SS designs out of the water, without dispute (e.g., Brystons).
I'm not sure that there can really be any significant 2nd harmonic distortion in my design (as a result of imbalances in the PP stage).
If switching to Mosfets reintroduces it, I can't see the advantage.
Nor am I convinced there would be a resultant significant DC offset,
as a mere result of 2nd harmonic distortion in the .05% range.
The addition of yet more CCS circuits seems to be superfluous and only muddy the waters,
and possibly the sound too.
Are you sure you can get equally good performance on all levels,
while introducing so many more parts and side-circuits?
Last edited:
If this type of circuit were to be substituted for my Mu-Follower design,
you would be actually re-introducing the non-linearity back into the circuit on each side.
Granted that some people would be actively seeking a little of this 'tube-distortion' (e.g., guitar-amp enthusiasts, and 'tube-sound' guys),
It would be mostly cancelled out again in the final Push-Pull stage of my original balanced design.
The residual distortion would comprise of odd-harmonics,
and other undesirable IM components.
Essentially, your idea would take a clean amp (dual SE or monoblock PP),
and attempt to add 2nd harmonic distortion, as a desirable feature.
This utterly defeats the zero-tolerance to distortion design of the original.
Those suggesting my amp would perform better by substitution of an SRPP, Anti-triode, or SRPP+ output stage,
have missed the entire point here,
by a subtle substitution of design goals.
There would be no advantage at all in doing this in the balanced Push-Pull monoblock version.
You would only add distortion which has already been carefully avoided, and then be imperfectly cancelling it out again in the primary of the transformer.
The result is poorer performance.
The misguided concern for 'current delivery' capability,
and impedance-matching based on unrelated circuits only ruins the design.
With complimentary Mu-Follower Push-Pull output stages,
you have the lowest possible distortion,
current delivery is addressed using other methods (like tube choice and operating point),
and impedance-matching is a red-herring.
The deliberate choice of a topology without DC-current flowing
through the primary coils cleans up the performance of any OT immensely,
and Output Impedance can be further adjusted by alternate taps and wiring of the secondary.
(or simply choosing an appropriate transformer).
I'm not sure the issue is Mosfet vs. Tube at all,
but rather SRPP+ vs. Mu-Follower here,
and what kind of philosophy you are adopting,
especially in regard to how you want SE power amps to perform.
I would advise that if you are going to deliberately introduce and manipulate 2nd Harmonic distortion,
you should do it in a preamp stage, where you can further control it with tone-controls and other performance enhancing circuitry.
The philosophy of the recording studio
is to do that at the mixing board and in mix-down,
not hope for it in the last stage of amplification, namely the power-amp.
- Status
- Not open for further replies.